This paper proposes a tree-based backoff （TBB） protocol that reduces the number of iterations implemented in the procedure of tag collision arbitration in radio frequency identification （RFID） systems. This is achieved by employing the following mechanisms： one is send the request command iteratively to all tags in the interrogation zone until a single tag is identified. The other is backward to the parent node instead of root node to obtain the request parameters and send the request command again until all tags are identified. Compared with the traditional tree-based protocol, on average, simulated results show that the TBB protocol reduces the number of the iterations by 72.3% and the identification delay by 58.6% and achieves the goal of fast tag identification.
分层次B帧是一种基于闭环运动补偿的时域可伸缩性视频编码,其特点是解码重构后的B帧可作为参考帧。针对分层次B帧的特点,提出一种新的B帧双向预测直接模式。通过挖掘相邻帧间的时域相关性,利用B帧作为参考帧时的前后向运动矢量进行时域缩放,精细当前块在直接模式下的运动矢量精度。仿真实验证明,该方法相对于传统的编码方式,编码性能增益平均提高0.46 d B。
